The 28 Śaiva Āgamas

In the great Śaiva temples of the Tamil country, the priest who bathes and dresses and feeds the god each morning is performing an act the Veda does not describe. The consecration that first made that stone a god, the syllables spoken over it, the geometry of its installation: these come from another scripture. The priest is an Ādiśaiva, and the book behind his hands is an Āgama.

There are said to be twenty-eight of them. The number is clean, closed, memorable. Ten revealed as Shiva, eighteen revealed as Rudra, and from these twenty-eight roots a further two hundred and seven subsidiary scriptures, so that the whole revelation of Shiva is held in a fixed and countable canon.

That is the tradition’s account, and it is worth stating in its own confident terms before complicating it, because the complication is the real subject. The number is a frame the tradition built. Behind the frame the texts are lost, partial, differently listed, and weighted almost entirely toward ritual. Both things are true at once. This is the canon of the Śaiva Siddhānta, one of the three great streams of the tantric revelation mapped in What are the Āgamas and Tantras?, and it is the scriptural backbone of temple Śaivism.

Ten from Shiva, eighteen from Rudra

The canon divides in two. Ten Śivabheda, the “Shiva-division” Āgamas, and eighteen Rudrabheda, the “Rudra-division.” Ten plus eighteen is twenty-eight.

The ten Śiva Āgamas run, in the most commonly cited list: Kāmika, Yogaja, Cintya, Kāraṇa, Ajita, Dīpta, Sūkṣma, Sahasra, Aṃśumat, Suprabheda. The eighteen Rudra Āgamas begin with the Vijaya and include the Niḥśvāsa, the Svāyambhuva, the Raurava, the Kiraṇa, the Mataṅga’s parent-text and the rest, ending by tradition with the Pārameśvara. The tradition ties the revelation to the five faces of Sadāśiva, the supreme form of Shiva from whose faces the scriptures are said to issue. The Aghora face, the fierce southern face, is said to give the Vijaya, the Niśvāsa, the Svāyambhuva and their group. The face-to-Āgama assignments vary by source, and that variation is the first honest signal that the list is a construction.

Because the list does not hold still. Dominic Goodall, editing the Kiraṇatantra commentary for the Institut Français de Pondichéry, devoted an appendix to the variant canon-lists, and the variation is real: names shift, orders change, totals wobble. Jean Filliozat’s own printed list of the classical Śaiva Āgamas runs to twenty-seven entries and merges the two groups. Most striking of all, two of the most philosophically central and best-preserved Śaiva scriptures, the Mṛgendra and the Mataṅgapārameśvara, are not among the standard twenty-eight at all. They are usually treated as upāgamas, subsidiary texts hanging off a root. The canon as a fixed list and the canon as a living authority do not quite line up.

The four quarters, and what survives

Each Āgama, in the ideal, carries four pādas, four quarters: jñāna, the doctrine; yoga, the discipline; kriyā, the ritual, which includes temple architecture and the making of images; and caryā, the conduct. Four departments of a complete religion in a single scripture.

The reality is lopsided. “Several of these texts seem to be lost,” Filliozat wrote flatly. “Many are restricted to one or two pādas instead of four. The priests ordinarily keep the kriyāpādas.” The ritual quarter survives because the ritual is used every day. The doctrinal and meditative quarters, less practically urgent, were more easily lost. What we hold of the canon is heavily its how, and more thinly its why.

Pati, pashu, and pasha

The why, where it survives, rests on three words. Pati, the Lord, Shiva. Paśu, the beast, the bound soul, the individual self in its fetters. Pāśa, the bond itself, the rope that ties the beast. These three, and their relations, are the metaphysical core of the Śaiva Siddhānta, set out most cleanly in the doctrinal sections of the Mṛgendra and the Mataṅga.

The bond, the pāśa, is threefold: āṇava, the primal impurity of finitude, the atomic smallness that makes a soul feel separate; karma, the residue of action; and māyā, the material stuff from which worlds are built. The three are held to be beginningless and real. This is a frankly dualist theology. God, souls, and the fetters are distinct, and liberation is not the soul dissolving into Shiva but the soul, freed of its bonds by grace and by initiation, coming to share Shiva’s nature while remaining itself. Richard Davis, whose Ritual in an Oscillating Universe reads the daily worship of Shiva as this theology enacted, calls the Siddhānta cosmos exactly that: oscillating, a universe Shiva emits and reabsorbs, in which the worshipper’s rite mirrors the pulse of creation.

There was internal argument over how dualist to be. In the twelfth century, Aghoraśiva of Chidambaram sharpened the dualism by pushing the five highest reality-levels, the tattvas from Nāda down to Śuddhavidyā, into the category of pāśa, of bond, a move earlier teachers had not made. His ritual manual, the Kriyākramadyotikā, remains the standard handbook of South Indian Śaiva worship to this day. A monistic strand persisted alongside him, and the frankly non-dual Śaivism of Kashmir stood outside the Siddhānta entirely. The article does not resolve the monist-dualist quarrel. It names it and lets both stand.

Where the Siddhānta sits

To see the twenty-eight clearly you have to see what they are not, and here Sanderson’s map of the whole Śaiva revelation is indispensable.

Sanderson calls the tantric Śaiva path the Mantramārga, the Way of Mantras, distinguishing it from an older, pre-tantric ascetic path, the Pāśupata Atimārga. The Mantramārga then divides. On one side stands the Śaiva Siddhānta: dualist, ritual-centred, worshipping Sadāśiva, socially respectable, the canon of the twenty-eight. On the other side stand the non-Saiddhāntika scriptures, the Bhairava tantras, more monistic, more Śākta, more transgressive, home to the Trika and the Krama and the goddess-centred cults that fed Kashmir Śaivism and the Śākta Tantras. The twenty-eight are one wing of a larger building. They are the dualist, temple-facing, publicly acceptable wing, which is exactly why they became the scripture of the great public temples.

The temple the Āgamas build

To call the Āgamas the scripture of the temples understates how literal that is. The kriyā quarter is, in large part, a construction manual. It lays down where a temple may stand and how it must be oriented, the vāstu of the site; the proportions and iconometry of the image, the śilpa by which a god is given a body in stone or bronze; the rite that brings that body to life, the pratiṣṭhā or kumbhābhiṣeka in which mantra and consecrated water install the divine presence, a rite the tradition holds should be renewed roughly every twelve years; and the daily round of worship that follows, in the larger temples a fixed cycle of services, as many as six a day. A Śaiva Āgama is, to a great degree, the operating system of a working temple.

These are not dead letters. The standard living handbook of South Indian Śaiva worship is the Kriyākramadyotikā of Aghoraśiva, completed in 1157, and the hereditary priests still perform from its descendants. The scholarship that recovered this world is recent and largely French. Hélène Brunner devoted four volumes to a single medieval ritual manual, the Somaśambhupaddhati, the first on the daily rite and the last on the consecration of temples. Bruno Dagens drew out the architectural teaching buried in the Ajita and Raurava Āgamas and translated the Mayamata, a treatise on building and image-making. Behind all of it stands the Institut Français de Pondichéry, which under Jean Filliozat gathered some eight thousand palm-leaf manuscripts, most in the southern Grantha script, and made the critical study of the Āgamas possible at all. The texts that looked lost were, for the most part, still in daily use. They had simply never been read as literature.

The priests, and initiation over birth

The Āgamas made a priesthood. The Ādiśaivas, also called Śivācāryas or, in Tamil, gurukkaḷ, are the hereditary officiants qualified to perform Āgamic temple ritual. Birth into the lineage matters.

But the deeper qualification, in the tradition’s own logic, is dīkṣā, initiation. The whole Śaiva Siddhānta turns on the claim that initiation transforms the soul, burns its bonds, makes it fit for liberation and for the handling of the god. Filliozat pressed the paradox to its edge: in the Śaiva reckoning the majority of Brahmins are not priests, and the priesthood is constituted as much by the transforming rite as by the accident of birth. A scripture that opens liberation through initiation carries, quietly, a different theory of who may approach the sacred than the Veda’s theory of birth.

North or south, and the shape of the dating

Two questions remain, and the honest answer to both is that the tradition and the record diverge.

The first is geography. Tradition, and much older scholarship, place the heart of the Śaiva Siddhānta in the Tamil South, and the living tradition is unmistakably Southern. But Sanderson and Goodall argue that the scriptural Siddhānta, a pan-Indian, Sanskrit, initiation-based religion, took shape earlier and very possibly elsewhere, in the North or in Kashmir, and only later became identified with the Tamil country. The Tamil devotional hymns, the Tirumurai of the saints, do not at their composition claim to belong to this scriptural canon. The Tamil face of the Siddhānta is, on this reading, a later and glorious identification, not the origin.

The second is date. The core Siddhānta Āgamas are placed by Sanderson roughly in the sixth to ninth centuries. The earliest named theologian we can fix is Sadyojyoti, around the seventh century, working already from the Raurava and the Svāyambhuva. Six of the classical Āgamas, the Pārameśvara, Niśvāsa, Svāyambhuva, Raurava, Kiraṇa and Parākhya, survive in copies demonstrably older than the twelfth century. Beyond those anchors the absolute dates of individual Āgamas are not fixed. “Sixth to ninth century” is a window for the formation of a canon, not a birth certificate for each text.

The Tamil systematisers

There is a last chapter to the canon, and it is written in Tamil. The Sanskrit Āgamas gave the Śaiva Siddhānta its ritual and the bones of its metaphysics. The philosophy, argued out and systematised, was the work of a later school in the Tamil country, four centuries and more after the core Āgamas took shape.

Its founding text is tiny and famous. The Śivajñānabodham, the “awakening to the knowledge of Shiva,” composed by Meykaṇṭār around 1221, is twelve terse sūtras on pati, paśu and pāśa and the soul’s road to union with Shiva. Tradition holds that the twelve render a passage of the Sanskrit Raurava Āgama, though the ascription is debated and some sources name the Pauṣkara instead. From this seed grew the Meykaṇṭa Cāttiram, the fourteen Tamil works of the thirteenth and fourteenth centuries that are the doctrinal core of the school. Meykaṇṭār’s pupil Aruḷnandi wrote the expansive Śivajñāna Siddhiyār, and Umāpati Śivācārya, in the early fourteenth century, added the luminous Śivaprakāśam and the Tiruvaruṭpayaṉ.

Their achievement was to turn the pati-paśu-pāśa theology into a rigorous pluralistic realism, in which God, souls and the bonds are all real and all beginningless, and liberation is the soul’s release, by grace, into the likeness of Shiva while remaining itself. This is the theology still taught in the great Tamil Śaiva monasteries, the ādhīnams of Dharmapuram and Tiruvāvaṭutuṟai, whose lineages reach back into this school.

The number was always a frame. What the frame held was a whole religion: a way to build a temple, install a god, cut a soul’s bonds, and think the matter through to the end.

There is a Tamil verse, cherished, that pictures the relation of the two revelations as a dairy: the Veda is the cow, the true Āgama its milk, the Tamil hymns of the saints its ghee. It is a beautiful image, and it is the tradition’s self-portrait, the Āgama as the drawn-out essence of the Veda. It is not a historical claim, and the article reports it as what it is: the way a tradition sees its own scripture, set beside, and kept distinct from, what the record will bear.

The tradition has kept these twenty-eight alive for fourteen centuries, most of them without ever counting the variants or worrying about the window of their dating. It did not need the count. But for a reader outside the temple, watching the priest’s hands and wondering where the words come from, the twenty-eight are the answer, held honestly: a real canon, a constructed number, a living authority, an imperfect list.